Control Your Game-Self: Effects of Controller Type on
Enjoyment, Motivation, and Personality in Game
Max Birk and Regan L. Mandryk
Department of Computer Science
University of Saskatchewan
110 Science Place, Saskatoon, SK, S7N 5C9
{firstname.lastname}@usask.ca
ABSTRACT
Whether they are made to entertain you, or to educate you,
good video games engage you. Significant research has
tried to understand engagement in games by measuring
player experience (PX). Traditionally, PX evaluation has
focused on the enjoyment of game, or the motivation of
players; these factors no doubt contribute to engagement,
but do decisions regarding play environment (e.g., the
choice of game controller) affect the player more deeply
than that? We apply self-determination theory (specifically
satisfaction of needs and self-discrepancy represented using
the five factors model of personality) to explain PX in an
experiment with controller type as the manipulation. Our
study shows that there are a number of effects of controller
on PX and in-game player personality. These findings
provide both a lens with which to view controller effects in
games and a guide for controller choice in the design of
new games. Our research demonstrates that including selfcharacteristics assessment in the PX evaluation toolbox is
valuable and useful for understanding player experience.
Author Keywords
Personality, motivation, self-discrepancy theory, selfdetermination theory, games, controller
ACM Classification Keywords
H.5.m. Information interfaces and presentation (e.g., HCI):
Miscellaneous.
General Terms
Human Factors; Design; Measurement.
INTRODUCTION
Whether they are made to entertain you, or to educate you,
good video games engage you. Good games draw the player
in and keep them engaged during play, and an engaged
player will give the game a better chance of success at its
intended purpose, irrespective of whether that be to divulge
or divert. Because it is so important for games to be
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. To copy otherwise,
or republish, to post on servers or to redistribute to lists, requires prior
specific permission and/or a fee.
CHI 2013, April 27–May 2, 2013, Paris, France.
Copyright © 2013 ACM 978-1-4503-1899-0/13/04...$15.00.
engaging, there has been significant effort from researchers
and practitioners in identifying and distilling the essential
ingredients for successful and engaging game design. These
efforts have taken both a prescriptive approach (i.e.,
attempting to formulate rules for game design), and an
analytic perspective (i.e., measuring play experience).
The analytic approach to understanding player experience
(PX) has taken many forms. Significant research has been
conducted on operationalizing concepts important to PX
(e.g., engagement, flow) [4], and measuring these concepts
by instrumenting [16], observing [13], and surveying [11]
the player. Although prior research on measuring PX
applies various methodological approaches, what previous
work has in common is that it all attempts to quantify or
qualify the player’s experience. However we suggest that to
fully understand PX, we must understand both how a player
feels about a game and how they feel about themselves
during gameplay. For example, consider a role-playing
game that rewards players who steal items from their
teammates. A player might be engaged and get satisfaction
from exercising their stealth inside the boundaries of the
game, but quit playing, because succeeding makes them
feel sleazy. In this case, the important information is not
how the player feels about the game, but about how the
game makes them feel about themselves.
Previous work has introduced the idea of considering a
user's experience of themselves during play by investigating
a user’s in-game trait personality; however, this application
of in-game personality has been only in the context of
sociological exploration of virtual worlds [39], association
with in-game avatars [2] or preference of game genre [23],
and not for the purpose of PX evaluation. To investigate the
relationship between player experience of a game and a
player’s in-game personality (how they felt about
themselves), we chose to compare the effects of different
game controllers. Studying game controllers is a natural
choice for this research as the controller acts as the interface
between the player and the game – it is the device through
which players engage with the game. Previous work
comparing input devices in gameplay has focused on
measuring performance differences in terms of targeting
[20] or steering [1], or player experience differences in
terms of spatial presence [34, 36] or immersion [25]. These
studies show that controller does affect PX, making
controller a good experimental manipulation – we assume
that different controllers will elicit different play
experiences; however, in this research we ask the question:
could the choice of game controller have an impact on the
player's perception of themselves during play?
We explore how the type of controller used during game
play affects the player's game experience in terms of their
enjoyment and their motivation. More importantly, we
show how controller choice affects a player's perception of
themselves and how their perceived personality (using the
five factor model [10]) within the game (game-self)
compares to their idealized version of themselves (idealself) and their reality (actual-self) [8]. To experimentally
determine the role of controller on PX, we designed and
implemented a custom game that allowed us to compare
three controller types (Microsoft Kinect, PlayStation Move,
and Xbox GamePad).
Our results show that the choice of controller affects our
enjoyment and motivation to play. Further, we show how
playing with different types of controllers changes how we
satisfy our needs within game play and how we perceive
ourselves in a game. Specifically, the choice of controller
affects our perceived Agreeableness and Neuroticism. In
addition, we show that change in game-self is the biggest
predictor of both enjoyment and motivation to play. These
findings provide both a lens with which to view controller
effects in games and a guide for controller choice in the
design of new games. Our research demonstrates that
including personality assessment in the PX evaluation
toolbox is valuable for understanding player experience.
RELATED WORK
In this section we provide an overview of existing
approaches to evaluating PX, the application of cognitive
evaluation theory and personality theory to understanding
PX, and previous work on how controller affects PX.
Evaluating Player Experience
Early attempts at game experience evaluation appropriated
methods from classic usability evaluation, but researchers
quickly discovered that games have different goals (e.g., to
challenge the user) and constraints (e.g., interrupting
players would break the experience) than productivity
software [9]. To combat these differences, researchers
proposed revising methods of traditional user experience
evaluation for use in games, including revised heuristics for
playability evaluation [26] and revised questionnaires
appropriate for the gaming context [27]. In addition, new
methods were created, including new survey instruments
(e.g., [11]), psychophysiological methods of evaluation
[16], and event log analysis [12] (what has now become
known as telemetry, game metrics, or game analytics [33]).
An increasingly active field of research, the goal of game
experience evaluation is to understand how the game affects
the emotional and cognitive experience of the player to
produce the desired outcome of ‘fun’ [5]. Although a few
evaluation approaches have been based on theory [37], the
grounding of PX evaluation in established theories of
human behaviour is lacking. Furthermore, a majority of
studies that measure game experience do not leverage the
research in validated evaluation approaches and instead
apply ad-hoc surveys to gather participant responses.
Self Determination Theory for PX Evaluation
The first psychological theory with validated scales that we
explore for understanding PX is Self Determination Theory
(SDT) [31]. We first describe the general theory and then
provide examples of its use in understanding PX.
Overview of SDT
SDT is comprised of several mini-theories and focuses on
factors that influence motivation. One mini theory,
Cognitive Evaluation Theory (CET) [31], describes
intrinsic motivation in terms of competence and autonomy.
To assess the factors that support intrinsic motivation in
games, the concept (and evaluation instrument) of Player
Experience of Needs Satisfaction (PENS) [32] was
developed. We next describe the concepts within PENS.
Intrinsic Motivation. In terms of SDT, intrinsic motivation
(IM) is defined by a locus of control inherent in the person,
and an outcome attributed to volition and achievement. IM
is distinct from external motivation, which is defined by an
external locus of control, e.g., deadlines. CET proposes that
IM is a function of autonomy, and competence.
Competence. CET proposes that the experience of
competence derives from challenge, and the personal effort
of mastering challenges. For example, the laddering used in
many games provides a structure for players to exert effort
to master successively harder challenges.
Autonomy. CET proposes that the experience of autonomy
derives from volition and willingness to perform a task. For
example, multiple in-game options give players cause to
experience autonomy through willing decision-making.
Relatedness. Within SDT, relatedness is defined as the
feeling of belonging to a group, and is an important factor
for psychological well-being. Social games like FarmVille
and massively multiplayer online role-playing games
(MMORPGs) make extensive use of this concept.
Presence. PENS defines presence as “the sense that one is
within the world”. Presence is fostered by competence and
autonomy, and related to the flow concept [4] as the highest
state of presence.
Intuitive Controls. PENS defines controls as intuitive,
when they do not interfere with one’s sense of presence, are
easily mastered, and make sense in context of the game.
Use of SDT in PX Evaluation
SDT has been applied in many psychological studies of
motivation (see [31]), but has recently been extended to
virtual environments and games. SDT was applied (using
PENS) to identify how different game genres satisfy needs
differently (e.g., experienced autonomy is highest in
roleplaying and strategy games) [11]. In the context of
game-based learning, [32] shows that experienced
competence, autonomy, and relatedness increase
motivation, positive mood, and the recommendation of a
game. Finally, in [24], the authors validate the application
of SDT (using PENS) to the domain of exergames, showing
that SDT-guided game feature choices result in improved
game enjoyment, game recommendation and game rating.
Specific to the evaluation of input, [17] applied PENS to
show that realistic and tangible controller mappings led to
higher experienced autonomy and presence.
play in several contexts. For example, the five factors of
personality have been shown to correspond with a player’s
preferred genre [23] and motivation to play [21]. In the area
of player taxonomies, research has shown that player type
can be predicted from the personality traits of the player
[19], and that considering personality traits in combination
with player taxonomies and game design elements can
inform the understanding of enjoyment in game play [29].
Additionally, research has shown that personality traits
carry into general behavior in virtual worlds [7].
Use of Self-Discrepancy Theory in Game Research
Self-Discrepancy Theory for Understanding Gameplay
The second psychological theory that we explore for
understanding PX is Self-Discrepancy Theory [8].
Although it has not been applied directly to PX evaluation,
it has been used to understand play. We first explain the
general theory and then give examples of its use in games.
Overview of Self-Discrepancy Theory
Self-discrepancy theory is based on the assumption that
people compare themselves to so called self-guides, or
internalized standards [8], and has been applied in fields
such as mental health and work-life balance (see [31]).
Self-guides represent three domains of the self, namely the
actual-self (i.e., the attributes we actually possess), the
ideal-self, (i.e., the attributes or characteristics we ideally
want to possess), and the ought-self (i.e., the attributes we –
or another person – believe we should have).
A common way to assess self-guides is to use the five
factor model of personality, commonly referenced as the
Big Five [10]. Usually used to assess personality traits, the
instructions are rephrased to use the following five factors
to assess different states of self-characterization.
Extraversion (E) is a tendency to be energetic, outgoing,
and assertive. People who score high on E tend to be
involved in social activities, rather than spend time alone.
Agreeableness (A) is a tendency to be friendly, caring, and
conflict avoiding. People who score high on A prefer cooperation to competition and are not likely to be aggressive.
Conscientiousness (C) is a tendency to show selfdiscipline, act dutifully, and aim for achievement. People
who score high on C plan activities rather than engaging in
spontaneity, and tend to be organized and dependable.
Neuroticism (N) is a tendency to be nervous, sensitive, and
emotionally unstable. People who score high on N respond
poorly to environmental stress, and are likely to experience
stress and frustration in minor situations.
Openness to Experience (O) is a tendency to be
intellectually curious, think abstractly, and to explore.
People who score high on O are likely to try something
new, rather than sticking to old beliefs and tradition.
Use of Personality Theory in Game Research
The use of personality theory in general – and the five
factors in particular – have been useful for understanding
In contrast to the direct application of personality traits
(e.g., Big Five factors), self-discrepancy theory considers
the difference between self-guides (e.g., the discrepancy
between actual-self and ideal-self), and has been applied
successfully to understand games. For example, research on
character choice in World of Warcraft has shown that
players create characters closer to their ideal-selves than to
their actual-selves [2], and that this trend was stronger for
those with higher depression scores. Additionally, research
has shown that the discrepancy between actual-self and
ideal-self is related to game enjoyment, especially when
convergence between game-self (self rated in the context of
game play) and ideal-self is high [28]. Finally, selfdiscrepancy has been used to explain game addiction [14].
Effects of Controller Choice on PX
Previous work comparing input devices in gameplay has
focused on measuring performance in terms of targeting
[20] or steering [1]. At the level of PX, research has shown
that movement-based input enhances engagement and
social interaction [15] in collaborative games. Others have
shown that controllers differ in terms of spatial presence
[36] and immersion [25] within games. It is important to
understand the spatial presence of controllers, because
research has shown that presence can predict enjoyment in
game play [34]. Finally, as noted in the section on the use
of SDT in PX evaluation, a realistic mapping between
controller and task leads to a higher experience of
autonomy and presence, but not competence [17].
EXPERIMENT DESCRIPTION
The goal of our study was to understand how playing a
game with different controllers alters our game experience
and our view on game-self characteristics. We compared
three controller types in a targeting game.
Game, System, and Apparatus
We designed a simple 3D game where the premise was that
the user was flying inside of a tornado in a ship and the goal
was to collect spinning items (Figure 1). Collection was
accomplished using a ‘tractor beam’ by hovering over the
item for 0.8 seconds. The beam changed colour to provide
feedback on successful hovering, and feedback on the
number of collected items were displayed in a gauge on the
ship’s console. Items appeared in waves of related items
(e.g., vegetables, construction items), and the player was
instructed to collect 10 items from each wave of 30 items.
Collecting 10 items allowed the player to speed in the ship
to the next wave. The number of waves in each game
depended on the success of the player and each game ended
after five minutes. Our design goal was to create a game
that was playable with a variety of controller types, while
keeping difficulty constant to allow inexperienced gamers –
or gamers inexperienced with a particular controller – the
opportunity to succeed in the task. Although complete in
terms of the formal elements of game design, our game was
specifically designed to be stripped of dramatic elements –
such as characters, story, and dramatic arc – to study how
controller choice affects self-perception and PX without
interference from dramatic effects, (such as an unfolding
narrative), or confounding effects, (such as how much the
player identifies with their in-game avatar) [2].
PlayStation Move button to start the game and the x- and yposition of the controller to position the tractor beam.
Kinect: The Microsoft Kinect is a gesture-based controller
developed for Xbox360. The device uses one camera that
detects x- and y-position and another infrared receiver
camera that detects depth through the dispersion of dots
displayed via an infrared transmitter – for a detailed
specification see [34]. In our study the experimenter started
the game; the x-position of the tractor beam was controlled
by the relative position of the left and right shoulder to the
Kinect; the y-position was controlled by the relative
position of hip and shoulder to the Kinect.
Training
Players completed a training condition consisting of aiming
at 8 targets positioned around the display using the same
selection mechanic as in the game (0.8-second dwell) prior
to each controller condition.
Participants and Procedure
The experiment was conducted with 78 students from the
University of Saskatchewan (38 female, mean age=25.8,
SD=6.6). Students were recruited via mailing lists and
notices on bulletin boards, and represented a variety of
disciplines of study. Most participants played games (94%):
73% played on computers, 44% on consoles, and 45% on
mobile devices. In terms of controller expertise, 51% were
experienced with the GamePad, 53% with the PlayStation
Move or Nintendo Wii, and 9% with Kinect1.
Figure 1. Screenshot of TornadoTwister.
The game was built using C# and XNA 4.0. We used the
Sony Move CL Eye Platform Driver and the Microsoft
Kinect SDK 1.5. The system ran on a Windows 7 PC with a
22” monitor with a resolution of 1680 by 1050. Players
stood 1.5 meters from the display in all conditions.
Controller Conditions
Our game required only that the player target objects by
manipulating the x- and y-position of the tractor beam;
there was no binary input necessary. We compared three
popular controller types (traditional, positional, and
gestural) represented by the Microsoft XBox GamePad, the
Sony PlayStation Move, and the Microsoft Kinect.
GamePad (GP): We used an Xbox360 standard controller.
For our study we used the A-Button to start the game and
the right analog stick to position the tractor beam.
Move (MO): The PlayStation Move is a position-based
controller, using a wand with a coloured light orb on top,
which is tracked by the PlayStation “Eye” Camera. The
PlayStation Eye was set to capture the scene at 60 Hz with a
resolution of 640 x 480, and a 76˚ field of view. The
location of the orb can be determined with an accuracy of
1mm and a latency of 22 ms. For our study, we used the
Participants began by completing an informed consent and
a questionnaire about their actual-self, ideal-self, and their
current affective state. They then played one game (exactly
5 minutes) with each of the three different controller types.
After each controller condition the players completed
questionnaires assessing their experience of the game and
of themselves within the game (game-self). At the end of
the experiment, players completed demographic questions.
The experiment was a within-subject design, with all
participants using each of the three controllers. Order of
presentation of controller was fully counterbalanced.
Measures
We collected all dependent measures using validated scales.
Cronbach’s-α is reported for PANAS, IMI, and PENS in
Table 1. For BFI, αactual-self =.641, and αideal-self =.669; αgameself is not reported because of the use of randomized
subscales; imputation would overestimate αgame-self.
1
Note that the novelty of the Kinect among our sample is a
potential confound. As such, we replicated our analyses for
a subgroup where players were unfamiliar with all devices
(N=31; 19 female). Our main results are all confirmed in
this subgroup, with two exceptions (flagged in the results),
where the trends of the results remain, but pMainEffect>.05.
Game Enjoyment
Enjoyment of game was measured using two scales.
PANAS: Positive Affect and Negative Affect were assessed
using the Positive Affect Negative Affect ScheduleExpanded (PANAS-X) [38]. In the PANAS-X, participants
are asked to agree with 20 emotion adjectives on a Likertscale ranging from 1 (very slightly or not at all) to 5
(extremely). Half of the adjectives are positive (e.g.,
‘active’) and half are negative (e.g., ‘guilty’). Ratings are
merged to create a composite score for negative affect and
one for positive affect. The PANAS-X has been used to
evaluate the enjoyment of video games (e.g., [28]).
IMI: Intrinsic Motivation was assessed using the 18-item
Intrinsic Motivation Inventory [30], which has been used to
evaluate experience with video games (e.g., [32]). A series
of items are rated on 5-point Likert-scale, ranging from 1
(not at all) to 5 (quite a bit). Example items are “I enjoyed
this game very much” and “I think I am pretty good at this
game”. Data is merged to create four scores for each of
interest-enjoyment, competence, effort-importance, and
tension, and also an overall score of intrinsic motivation.
Need Satisfaction
The Player Experience of Need Satisfaction Scale (PENS)
[32], was designed to understand game experience from the
perspective of SDT [31], and has been used to evaluate
games (e.g., [32]). We used PENS after each condition to
assess if the game satisfied the players’ needs for
Competence, Autonomy, Relatedness, Immersion and
Intuitive Controls. A series of statements was agreed with
using a 5-point Likert-scale from 1 (not at all) to 5 (quite a
bit). Example items include “I feel competent at the game”
or “Learning the game controls was easy”. Data are merged
to create a score for each of the five underlying constructs.
Although some have argued for leaving out the subscales
on Relatedness if the game is for a single player [11], it is
possible that different controller types could differentially
satisfy the need for relatedness, so we included it.
Personality
The player’s trait personality was assessed using the Big
Five Inventory (BFI) [10]. The BFI is a standard instrument
to assess personality dimensions and has been used in the
context of games [6]. Data are merged to create a score for
each of the underlying dimensions and for an overall score.
Self-Discrepancy Theory assumes that the distance between
self-concepts is predictive for psychological well-being.
There are many ways to measure differences between selfconcepts, including personality questionnaires, which easily
assess personality constructs and get a measure for SelfDiscrepancy at the same time. We use the overall score
from the BFI to give a simple estimate for Self-Discrepancy
by calculating divergence (see [28] for further explanation).
Participants completed the BFI five times. At the beginning
of the experiment, participants were asked to reflect on their
actual-self and rated their agreement with 44 statements of
the form, “The type of person you are…”, with answers
such as “is talkative” or “likes to cooperate with others”.
Ratings used a 5-point Likert scale from 1 (strongly
disagree) to 5 (strongly agree). Participants also reflected
on their ideal-self, by completing items of the form “The
type of person you wish, desire, or hope to be…”, with the
same items as used for assessing actual-self. After each
game condition, participants were asked to assess their
game-self using a short form (15 questions). Participants
were asked to answer the questionnaire by completing the
sentence “Reflect on the characteristics you see yourself
having during the game – the type of person you were
playing the game”. Items in the short version covered all 5
aspects of personality, the number of items representing the
scales E, A, O, N, C per questionnaire were stable, and we
fully counterbalanced the presentation. Drawing items from
a longer questionnaire is an established approach [39].
Data Analyses
All analyses were conducted using RM-ANOVA in SPSS
20. All parametric tests were performed after validating the
data for assumptions of ANOVA use. Degrees of freedom
were corrected using the Huynh-Feldt method, if the
condition of sphericity was not satisfied. Pairwise
comparisons used the Bonferonni method of adjusting the
degrees of freedom for multiple comparisons, with the
exception of the self-discrepancy measures, which used
Tukey’s LSD to avoid potential Type II errors resulting
from small effect sizes. Significance was set at α=.05.
RESULTS
We first present the results of PX as traditionally measured
by positive affect and intrinsic motivation. We extend this
traditional analysis of PX by presenting the results of need
satisfaction. We then explore how the controller conditions
changed how players feel about themselves within the
context of the game through presentation of the personality
results. Finally, we investigate how the use of personality as
a measure of player experience compares with the use of
measures of play experience through a regression analysis.
Affect and Intrinsic Motivation
When evaluating PX, the first question of interest is usually
whether or not the experience was fun, often measured as
the degree of positive affect (e.g., [28]), mediated by the
degree of Intrinsic Motivation (e.g., [32]) – especially the
motivational component of interest/enjoyment.
Our results show a significant main effect of controller on
positive affect (F2,154=29.1, p≈.000, η2=.27). Pairwise
comparisons show that the Kinect produced the highest
Positive Affect, then the Move then GamePad (see Figure
2). All differences were significant at p<.006. There was no
controller effect on Negative Affect (F1.9,145.5=2.4, p=.101).
When considered as a whole, there was no effect of
controller on overall Intrinsic Motivation (F2,154=1.1,
p=.352); however, controller did affect the InterestEnjoyment subscale (F2,154=9.4, p=.001, η²=0.10) and the
Competence subscale (F1.9,144.2=8.9, p≈.000, η²=0.10).
Pairwise comparisons show that the GamePad was less
enjoyable than the Move (p=.031) or Kinect (p=.001),
which were not different from each other (p=.148). For
competence, the Move produced greater confidence than
the Kinect (p≈.000) or the GamePad (p=.002), which were
not different from each other (p=1.0). See Figure 3. There
were no main effects of controller on the Effort (F2,154=1.7,
p=.183) or Tension (F2,154=1.9, p=.147) subscales.
Relatedness than GamePad (p=.014); the other differences
were not significant (MO-KI: p=.281; MO-GP: p=.732).
There was a main effect of controller on Immersion
(F2,154=5.9, p=.003, η2=0.07). The Kinect provided higher
Immersion than GamePad (p=.004); the other differences
were not significant (MO-KI: p=.240; MO-GP: p=.272).
There was a main effect of controller on Intuitive Controls
(F1.8,137.0=25.2, p≈.000, η2=0.25). The Move was most
Intuitive, followed by the GamePad, then Kinect. All
differences were significant (MO-GP: p≈.000; MO-KI:
p≈.000; GP-KI: p=.025).
GamePad
M
SD
Figure 2. Means (±SE) for positive affect and negative affect
on a scale of 1 (low) to 5 (high).
Figure 3. Means (±SE) for subscales of intrinsic motivation on
a scale of 1 (low) to 5 (high).
Player Experience of Need Satisfaction
Although the analysis of affect and motivation provide
insight into the game experience provided by different
controllers, we used PENS, as it has been shown to provide
a deeper understanding of game experience [32]. See Table
1 for descriptive statistics, and Figure 4.
Figure 4. Means (±SE) for subscales of PENS on a scale of 1
(low) to 5 (high).
There was a main effect of controller on Competence
(F1.9,145.2=11.4, p≈.000, η2=0.13). Pairwise comparisons
revealed that the Move provided higher Competence than
the GamePad (p=.001) or Kinect (p≈.000), which were not
different from each other (p=.702).
There was a main effect of controller on Autonomy
(F2,154=9.3, p≈.000, η2=0.11). The Kinect provided higher
Autonomy than the Move (p=.044) and GamePad (p=.001),
which were not different from each other (p=.119).
There was a main effect of controller on Relatedness1
(F2,154=4.3, p=.016, η2=0.05). The Kinect provided higher
IMI
Combined
Interest-Enj.
Competence
Effort-Imp.
Tension
Affect
Positive
Negative
PENS
Competence
Autonomy
Relatedness
Immersion
Intuitive Cont.
Move
M
SD
M
Kinect
SD
α
3.15
2.96
3.08
3.66
2.90
0.52
0.86
1.05
0.91
1.12
3.23
3.19
3.51
3.55
2.66
0.60
0.93
0.92
1.00
1.04
3.22
3.37
2.97
3.71
2.83
0.60
1.06
1.04
0.86
1.09
0.82
0.90
0.83
0.86
2.76
1.52
0.89
0.67
2.98
1.41
0.93
0.51
3.33
1.46
0.91
0.55
0.90
0.86
3.15
2.31
2.15
2.32
3.75
1.02
1.14
0.93
0.91
1.06
3.56
2.49
2.24
2.44
4.29
0.91
1.13
0.92
0.97
0.69
2.97
2.70
2.36
2.56
3.32
1.06
1.22
1.06
1.08
1.11
0.83
0.85
0.67
0.89
0.72
Table 1. Means and SD for intrinsic motivation, affect and
PENS on a scale from 1 (low) to 5 high. Cronbach’s-α shows
the consistency of the scale items for each construct.
Perceived in-game Personality
The results for self-characteristics are presented overall and
for the five subscales (see Table 2 for descriptive statistics).
CS
E
A
C
N
O
Actual-self
M
SD
3.41 0.26
3.27 0.72
3.78 0.54
3.53 0.64
2.78 0.71
3.70 0.53
Ideal-self
M
SD
3.80 0.19
4.16 0.46
4.45 0.46
4.66 0.29
1.39 0.34
4.34 0.40
GamePad
M
SD
3.16
2.94
3.49
3.49
2.68
3.20
0.37
0.82
0.67
0.69
1.02
0.70
Move
M
SD
3.26
3.15
3.77
3.55
2.37
3.44
0.39
0.87
0.72
0.71
0.83
0.68
Kinect
M
SD
3.14
2.97
3.50
3.50
2.42
3.30
0.43
0.84
0.72
0.81
0.84
0.76
Table 2. Means and SD for personality by context (CS:
composite score, E: Extraversion, A: Agreeableness, C:
Conscientiousness, N: Neuroticism, O: Openness).
There was a main effect of controller on overall in-game
personality (F2,78=3.06, p=.050, η2 =0.04). The Kinect was
higher than the GamePad (p=.050) or the Move (p=.021),
which were not different from each other (p=.708).
Subscale analysis revealed a main effect of controller on
Agreeableness (F1.9,145.3=6.05, p=.004, η2=0.07) and
Neuroticism1 (F2,154=3.4, p=.038, η2=0.04), but not on
Extroversion (F2,154=2.4, p=.100), Conscientiousness
(F2,154=0.2, p=.795), or Openness (F2,154=2.7, p=.071).
Pairwise comparisons show that the Kinect yielded more
Agreeableness than both the Move (p=.004) and GamePad
(p=.001), which were not different from each other
(p=.915). Experienced Neuroticism was higher for the
GamePad than the Move (p=.047) or Kinect (p=.031),
which were not different from each other (p=.687).
Immersion. For the Move condition, the correlation is
significant with Competence and Relatedness, and for the
Kinect condition, the correlation is significant with
Competence, Autonomy, and Relatedness. These results
reveal that PX is not related to players’ overall personality
traits, but is related to their experienced personality within
the context of game play, demonstrating that the use of
game-self-characteristics for PX is meaningful.
Com.
Aut.
Rel.
Imm.
Int.
Actual-self
r
p
.047
.479
-.054 .415
-.014 .837
-.069 .296
-.031 .635
GS: GamePad
r
p
.170
.070
.271
.008
.314
.003
.233
.020
.150
.097
GS: Move
r
p
.294
.005
.130
.121
.272
.008
.140
.106
.050
.319
GS: Kinect
r
p
.295
.004
.276
.007
.340
.001
.191
.047
.150
.088
Table 3. r and p values for correlations between PENS
dimensions and personality characteristics.
Added Value of Personality Analysis
Figure 5. Personality dimension means for: actual-self, idealself, and the game-selves (GP, MO, KI) of 1 (low) to 5 (high).
Figure 5 shows the differences in perceived personality in
game for the different controllers in the context of
perceived actual-self and ideal-self. Interestingly, the ingame personalities are contained within the boundaries of
actual-self. Contrary to expectations of game play, the game
did not produce more idealized versions of players, but
moved their perceived personality further from their ideal
than their actual-self (with the exception of neuroticism).
This ideal-self-game-self discrepancy is significantly larger
for all controller types than for ideal-self-actual-self (all
p<.001). We revisit this interesting result in the discussion.
Personality and Need Satisfaction
Figure 5 shows how in-game personality characteristics
change with controller type, but also shows how game-self
is fairly similar to actual-self. To determine whether the
differences in game-self are meaningful for game
evaluation, we conducted two-tailed Pearson correlations
for game-self and the PENS dimensions separately for each
controller. We also calculated the correlations for actualself and the PENS dimensions across the three controllers.
Correlations are displayed in Table 3.
The results show that there are no significant correlations
between actual-self and any of the PENS dimensions;
however, there are correlations between the PENS
dimensions and game-self. For the GamePad condition, the
correlation is significant with Autonomy, Relatedness, and
The results of the personality analysis show that a choice as
simple as which controller is used can affect our in-game
personality, and that in-game personality correlates with
need satisfaction, whereas actual-self does not; however to
understand the role that in-game personality plays in game
experience evaluation, we must know the relative
contribution of the need satisfaction variables and the
personality variables to the overall outcome of increased
positive affect or motivation to play. We conducted an
analysis of regression with game-self, Competence,
Autonomy, Relatedness, Immersion, and Intuitive Control
on Positive Affect and overall Intrinsic Motivation (see
Table 4 for standardized β-coefficients and p-values).
Personality was included as a single factor, indicating
overall game-self, whereas the individual contributions
from PENS were included individually because they
measure distinct dimensions; there is no meaning of an
overall PENS score.
Game-self
Competence
Autonomy
Relatedness
Immersion
Intuitive Control
Positive Affect
β
P
0.31
.000
0.23
.001
0.28
.000
0.05
.538
0.09
.296
-0.06
.396
Intrinsic Motivation
β
p
0.18
.000
0.11
.100
0.11
.128
-0.01
.918
0.46
.000
0.12
.057
Table 4. Standardized beta coefficients and p values for
regression of game-self and PENS dimensions on positive
Our results show that game-self, Competence, and
Autonomy predict Positive Affect, with changes in gameself yielding the biggest contribution. In addition, gameself and Immersion predict Intrinsic Motivation, with
Immersion providing the larger contribution. These results
show that changes to game-self as a result of controller
predicts positive affect and intrinsic motivation during
game play, and confirm that analysis of in-game personality
is a valuable tool to understand PX.
DISCUSSION
Our study shows that there are a number of effects of
controller on PX and in-game personality. We describe the
significance of the findings, discuss the implications for
game design, and finally present the limitations of our study
and the future opportunities that our findings provide.
Significance of the Findings
Other researchers have previously applied theories of needs
satisfaction to understand and explain player motivation
and user experience within a game (e.g., [11, 17, 32]). We
do the same in our work but add to the literature with our
findings on how controller choice differentially affects our
satisfaction of needs. In addition, we demonstrate that the
application of self-discrepancy theory to game evaluation
adds value to our understanding of PX. Game-self changed
significantly with use of a different controller, and the
change in game-self as a result of controller choice is a
bigger predictor of positive affect than the components of
need satisfaction theory. These findings confirm that
including trait personality assessment in the PX evaluation
toolbox is valuable for understanding player experience.
Our findings also have practical importance. For instance,
we use games to escape reality, whether through a fantasybased role-playing experience (e.g., World of Warcraft) or
not (e.g., Bejeweled). We show that controller choice
affects game-self, thus it is important for game designers to
consider controller choice, to ensure that the escapism
provided by games is not unintentionally damaged.
Implications for Game Design
Although the primary purpose of computer and video
games has been to entertain, games have also been used to
persuade a player to change their attitudes or opinions (e.g.,
The Cat and the Coup, A Closed World), to inspire
behaviour change (e.g., Evoke, Fatworld), and to teach
concepts, skills and tools (e.g., Math Blaster, Typing of the
Dead, RibbonHero) [18, 3]. Regardless of whether a game
is intended to entertain or educate, players will enjoy the
experience of play if they are motivated to make progress
[32]. Good games draw the player in and keep them
engaged during play, and an engaged player will give the
game a better chance of success at its intended purpose,
irrespective of whether that be to divulge or divert.
Our results demonstrate how the choice of controller in a
game mediates the experience, not only for the satisfaction
of needs, but also for perceived personality. These findings
related to game-self and controllers are useful both as a lens
with which to understand controller effects in games and as
a guide for choices on controller in new games.
Findings as a Lens to View Existing Games
Our findings on controller and game-self could be used to
understand player experience in a variety of games and
situations. For example, consider games where the player
must help the character overcome obstacles and prevail
(e.g., Deus Ex series, Mass Effect series). In these games,
the character begins with vulnerabilities (e.g., low skills)
that are resolved through interacting with environmental
threats. The designer’s intention is that players experience
the fragility of the hero and surmount the odds; the goal is
to become strong and overcome the feelings of being weak.
Players accomplish this through leveling up (which happens
to also satisfy our need for Competence). Many of these
types of games use traditional controllers (e.g., GamePad),
which is shown in our study to increase Neuroticism. By
using the GamePad for these types of interactions, the
experience of anxiety and instability might be heightened;
users may be more likely to experience stress from minor
events. It could be that in playing these games with a
GamePad, the designer’s intention of having players
experience the character’s fragility is effectively translated.
In another example, consider games designed as social
icebreakers or to promote social connectedness (e.g., party
games such as Wario’s Smooth Moves). Games that target a
friendly play environment benefit when players’ gameselves are Agreeable. Our results suggest that the Kinect is
preferred for enhancing Agreeableness, matching what is
working in the games industry; the success of the Kinect
has largely been for these types of social games.
Findings as a Guide to Inform New Game Design
Our findings related to in-game personality as a result of
controller choice could also be applied to guide controller
choice in the design of new games.
For example, consider a game designed to help players
learn empathy. Our findings show that using the GamePad
heightens the experience of Neuroticism, reinforcing
feelings of being uncomfortable with strangers and difficult
social situations, and causing players’ game selves to be
more likely to lose their temper and be prone to annoyance.
Our results suggest that using the GamePad would decrease
the game-self’s natural inclination towards empathy, thus
increasing the challenge of experiencing empathy;
therefore, it could be a good choice for this application.
In another example, consider a game that encourages
deception and betraying others for personal gain (e.g.,
DayZ). Although the Kinect is a general choice for
collaborative or social games, our results show that the
Kinect enhances players’ Agreeableness within the context
of play, reducing willingness to take advantage of others,
insult others, and invite conflict. Thus, our findings suggest
that the Kinect may not be a good controller choice when
designing a game where the mechanics incite low
Agreeableness (e.g., by focusing on personal advantage and
conflict). The mismatch between the personality
affordances of the controller and the intended in-game
personality of the player may result in poor PX.
On the other hand, as previously noted, our findings suggest
that games designed as social icebreakers or to promote
social connectedness would benefit from the Kinect.
Opportunities for Future Work
Our study provides interesting findings for PX research;
however, there are some limitations that raise questions for
future work. First, our game was intentionally designed to
be void of dramatic elements related to story arc or
character progression. How our results extend to games
where the player relates to their character is unknown. Our
results that show how game-self is subsumed by actual-self
likely do not extend to situations in which a player interacts
via a character with personality traits closer to the player’s
ideal-self [2]. Second, although our game was engaging, it
was not designed to draw a player in like an off-the-shelf
game. Our results may differ when applied to games
engineered to produce engagement and flow. On the other
hand, the underlying mechanism of pursuit tracking used in
our game is a common mechanic in a variety of game
genres (e.g., first-person shooter), and thus our results
related to controller choice have general application. Third,
we tested three controllers; extending our results to other
game input (e.g., mice, joysticks, touch) is needed. Also, we
created a specific control mechanism for Kinect interaction
(using torso movement); a different control mechanism
(e.g., hand position) may produce different effects on gameself. Finally, our analysis considers all participants as a
group. Examining results by demographic factor, such as
age, sex, personality traits, or game expertise could produce
interesting interactions.
Our findings produce many future opportunities for
research, of which we present a few examples. First, we
investigated a single-player game; applying our approach to
collaborative or competitive play could inform the choice of
controller for this popular genre. Second, we look
specifically at controller choice in digital games; our
analysis could be extended to artifact use in board game
play. Finally, our findings can be used as a lens to view
existing game success. Relating the predictions of positive
affect based on changing game-self to market success of
games would provide commerce-based validation of our
research.
CONCLUSIONS
Traditional game evaluation focuses on the enjoyment
provided by a game (i.e., positive affect). We apply
satisfaction of needs theory (using PENS) to explain the
underlying components of game enjoyment in an
experiment with controller type as the manipulation. In
addition, we explore the use of game-self as an additional
level of PX evaluation to investigate the difference between
how a play experience makes a player feel about the game
and how the experience makes them feel about themselves.
Our study shows that there are a number of effects of
controller on PX and player personality within a game.
These findings provide both a lens with which to view
controller effects in games and a guide for controller choice
in the design of new games. Finally, our research
demonstrates that including self-characteristics assessment
in the PX evaluation toolbox is valuable and useful for
understanding player experience.
ACKNOWLEDGEMENTS
Thanks to reviewers for their feedback, the GRAND NCE
for funding, and Kristen Dergousoff, Roxanne Dowd, Jared
Cechanowicz, and the rest of the Interaction Lab at the
UofS for support.
REFERENCES
1.Bateman, S., Doucette, A., Xiao, R.B., Gutwin, C.,
Mandryk, R.L., & Cockburn, A. Effects of View, Input
Device, and Track Width on Video Game Driving. In
Proc. GI, 2011, 207-214.
2.Bessiere, K., Seay, & F. Kiesler, S., The Ideal Elf:
Identity Exploration in
World of
Warcraft.
CyberPsychology & Behavior, 2007, 10(4), 530-535.
3.Bogost, I., How To Do Things With Videogames, 2011.
4.Csíkszentmihályi, M., Beyond Boredom and Anxiety,
1975.
5.Fernandez, A., Fun Experience with Digital Games: A
Model Proposition. In Extending Experiences: Structure,
Analysis and Design of Computer Game Player
Experience, 2008, 181-190.
6.Gilbert, R. L., Foss J.A., & Murphy, N.A., Multi
Personality
Order:
Physical
and
Personality
Characteristics of the Self, Primary Avatar and Alt. In
Reinventing Ourselves: Contemporary Concepts of
Identity in Virtual Worlds, 2011, 213-234.
7.Hefner, D., Klimmt, C. and Vorderer, P., Identification
with the player character as determinant of video game
enjoyment. In Proc. of ICEC, 2007, 39-48.
8.Higgins, E.T., Self-Discrepancy: A theory relating self
and affect. Psychological Review, 1987, 94(3), 319-340.
9.Isbister, K., & Schaffer, N., Game Usability: Advice from
the Experts for Advancing the Player Experience, 2008.
10.John, O. P., Naumann, L. P., & Soto, C. J., Paradigm
Shift to the Integrative Big-Five Trait Taxonomy: History,
Measurement, and Conceptual Issues. In Handbook of
personality: Theory and research, 2008, 114-158.
11.Johnson, D., & Gardner, J., Personality, Motivation and
Video Games (2010). In Proc. of OZCHI, 2010, 276-279.
12.Kim, J., Gunn, D., Schuh, E., Phillips, B., Pagulayan, R.,
& Wixon, D. Tracking Real-Time User Experience
(TRUE): A comprehensive instrumentation solution for
complex systems. In Proc. of SIGCHI, 2008, 443-452.
13.Lazzaro, N., Why we play games: 4 keys to more
emotion. In Proc. of GDC, 2004.
14.Li D., Liau A., & Khoo A., Examining the Influence of
Actual-Ideal-Self-Discrepancies,
Depression,
and
Escapism, on Pathological Gaming Among Massively
Multiplayer Online Adolescent Gamers Cyberpsychology,
Behavior, and Social Networking, 2011, 14(9), 535-539.
15.Lindley S. E., Le Couteur J., & Berthouze N. L, Stirring
up experience through movement in game play: effects on
engagement and social behaviour. In Proc. CHI, 2008,
511-514.
16.Mandryk, R.L., Atkins, M., & Inkpen, K., A Continuous
and Objective Evaluation of Emotional Experience with
Interactive Play Environments. In Proc. of CHI, 2006,
1027-1036.
17.McEwan, M., Johnson, D., Wyeth, P. & Blacker, A.
Videogame Control Device Impact on the Play
Experience. In Proc. of IE, 2012, 1-3.
18.McGonigal, J., Reality Is Broken: Why Games Make Us
Better and How They Can Change the World, 2010.
19.McMahon, N., Wyeth, P. & Johnson, D., Personality and
Player Types in Fallout New Vegas. In Proc. of FNG,
2012, 113-116.
20.Napatov, D.,Castellucci, S.J., & MacKenzie, I.S., ISO
9241-9 evaluation of video game controllers. In Proc. of
GI, 2009, 223-230.
21.Park J., Song Y., & Teng C., Exploring the Links
Between Personality Traits and Motivations to Play
Online Games. Cyberpsychology, Behavior, and Social
Networking, 2011, 14(12) 747-751.
22.Partala, T., Psychological needs and virtual worlds: Case
Second Life, International Journal of Human-Computer
Studies, 2011, 69(12), 787-800.
23.Peever, N., Johnson, D., & Gardner, J., Personality and
Video Game Genre Preferences. In Proc.of IE, 2012
24.Peng, W., Lin, J., Pfeiffer. P., & Winn, B., Need
Satisfaction Supportive Game Features as Motivational
Determinants: An Experimental Study of a SelfDetermination Theory Guided Exergame, Media
Psychology, 2012, 15(2), 175-196.
25.Pietschmann D., Valtin, G., & Ohler O., The effect of
Authentic Input Devices on Computer Game Immersion.
In Computer Games and New Media Cultures: A
Handbook of Digital Games Studies, 2012.
26. Pinelle, D., Wong, N., Stach, T., & Gutwin, C.,
Usability Heuristics for Networked Multiplayer Games,
In Proc. of GROUP, 2009, 169-178.
27.Procci, K., Singer, A. R., Levy, K. R., & Bowers, C.,
Measuring the flow experience of gamers: An evaluation
of the DFS-2. Computers in Human Behavior, 2012,
28(6), 2306-2312.
28.Przybylski, A.K., Weinstein, N., Murayama, K., Lynch,
M.F., & Ryan, R.M., The ideal-self at play: The appeal of
video games that let you be all you can be. Psychological
Science, 2011, 23 (1), 69-76.
29.Quick, J. M., Atkinson, R. K., & Lin, L., Empirical
Taxonomies of Gameplay Enjoyment: Personality and
Video Game Preference. International Journal of GameBased Learning, 2012, 2(3), 11-31.
30.Ryan, R. M., Control and information in the
intrapersonal sphere: An extension of cognitive evaluation
theory. Journal of Personality and Social Psychology,
1982, 43, 450-461.
31.Ryan, R. M., & Deci, E. L., Self-determination theory
and the facilitation of intrinsic motivation, social
development, and well-being. American Psychologist,
2000b, 55, 68-78.
32.Ryan, R. M., Rigby, C. S., & Przybylski, A. K.,
Motivational pull of video games: A self-determination
theory approach. Motivation and Emotion, 2006, 30, 347365.
33.Seif El-Nasr, M., Drachen A., Canossa A., Game
Analytics: Maximizing the Value of Player Data.
Springer, (to appear).
34.Shafer, D.M., Carbonara, C.P., & Popova, L., Spatial
Presence and Perceived Reality as Predictors of MotionBased Video Game Enjoyment, Presence, 2011, 20(6),
591-619.
35.Sheldon, K. M., & Filak, V., Manipulating autonomy,
competence and relatedness support in a game-learning
context: New evidence that all three needs matter. British
Journal of Social Psychology, 2008, 47, 267-283.
36.Skalski, P., Tamborini, R., Shelton, A., Buncher, M. &
Lindmark, P., Mapping the road to fun: natural video
game controllers, presence, and game enjoyment. New
Media & Society, 2010, 224-242.
37.Sweetser, P. & Wyeth, P., GameFlow: a model for
evaluating player enjoyment in games. Computer
Entertainment, 2005, 3-3.
38.Watson, D., & Clark, L. A. The PANAS-X: Manual for
the positive and negative affect schedule-Expanded Form,
1994.
39.Yee. N., Ducheneaut, N., Nelson, L., & Likarish, P.
Introverted Elves & Conscientious Gnomes: The
Expression of Personality in World of Warcraft. In Proc.
of CHI, 2011, 753-762.